2.5.70 Adaptec error messages

2.5.70 Adaptec error messages

[Margit Schubert-While]

With 2.5.70(-bk) and also Justin's 20030603, I am getting the below at boot.
After that, everything works fine.
Adapter is 39160.

Margit

----------------------------------------
kernel: (scsi0:A:0:0): CDB: 0x12 0x1 0x80 0x0 0xff 0x0
kernel: (scsi0:A:0:0): Saw underflow (752 of 768 bytes). Treated as error
         Repeats another 4 times
kernel: (scsi0:A:1:0): CDB: 0x12 0x1 0x80 0x0 0xff 0x0
kernel: (scsi0:A:1:0): Saw underflow (752 of 768 bytes). Treated as error
         Repeats another 4 times
kernel: (scsi1:A:2:0): CDB: 0x12 0x1 0x80 0x0 0xff 0x0
kernel: (scsi1:A:2:0): Saw underflow (752 of 768 bytes). Treated as error
         Repeats another 4 times
kernel: (scsi1:A:3:0): CDB: 0x12 0x1 0x80 0x0 0xff 0x0
kernel: (scsi1:A:3:0): Saw underflow (752 of 768 bytes). Treated as error
         Repeats another 4 times
----------------------------------------
         And the above block repeats once more

-- snip --
<6>scsi0 : Adaptec AIC7XXX EISA/VLB/PCI SCSI HBA DRIVER, Rev 6.2.36
<4>        <Adaptec 3960D Ultra160 SCSI adapter>
<4>        aic7899: Ultra160 Wide Channel A, SCSI Id=7, 32/253 SCBs
<4>
<4>(scsi0:A:0): 160.000MB/s transfers (80.000MHz DT, offset 127, 16bit)
<4>(scsi0:A:1): 160.000MB/s transfers (80.000MHz DT, offset 127, 16bit)
<5>  Vendor: FUJITSU   Model: MAN3184MP         Rev: 0108
<5>  Type:   Direct-Access                      ANSI SCSI revision: 03
<4>scsi0:A:0:0: Tagged Queuing enabled.  Depth 32
<5>  Vendor: FUJITSU   Model: MAN3184MP         Rev: 0108
<5>  Type:   Direct-Access                      ANSI SCSI revision: 03
<4>scsi0:A:1:0: Tagged Queuing enabled.  Depth 32
<6>scsi1 : Adaptec AIC7XXX EISA/VLB/PCI SCSI HBA DRIVER, Rev 6.2.36
<4>        <Adaptec 3960D Ultra160 SCSI adapter>
<4>        aic7899: Ultra160 Wide Channel B, SCSI Id=7, 32/253 SCBs
<4>
<4>(scsi1:A:2): 160.000MB/s transfers (80.000MHz DT, offset 63, 16bit)
<4>(scsi1:A:3): 160.000MB/s transfers (80.000MHz DT, offset 63, 16bit)
<5>  Vendor: IBM       Model: DDYS-T18350N      Rev: S96H
<5>  Type:   Direct-Access                      ANSI SCSI revision: 03
<4>scsi1:A:2:0: Tagged Queuing enabled.  Depth 32
<5>  Vendor: IBM       Model: DDYS-T18350N      Rev: S96H
<5>  Type:   Direct-Access                      ANSI SCSI revision: 03
<4>scsi1:A:3:0: Tagged Queuing enabled.  Depth 32
<6>scsi2 : Adaptec AIC7XXX EISA/VLB/PCI SCSI HBA DRIVER, Rev 6.2.36
<4>        <Adaptec 2940A Ultra SCSI adapter>
<4>        aic7860: Ultra Single Channel A, SCSI Id=7, 3/253 SCBs
<4>
-- end snip --

Re: 2.5.70 Adaptec error messages

[Justin T. Gibbs]

> With 2.5.70(-bk) and also Justin's 20030603, I am getting the below at boot.
> After that, everything works fine.
> Adapter is 39160.

Someone is fetching the device serial number without setting the
cmd underflow to 0.  The response length is variable, so telling
the adapter to fail the command when the returned result is less
than the allocated buffer size is counter productive.

--
Justin

Re: 2.5.70 Adaptec error messages

[Margit Schubert-While]

At 10:01 05.06.2003 -0600, Justin wrote:
>Someone is fetching the device serial number without setting the
>cmd underflow to 0.  The response length is variable, so telling
>the adapter to fail the command when the returned result is less
>than the allocated buffer size is counter productive.

Hmm, doesn't happen with 2.4.21rc + Justin's driver.
Which, for me, would imply the kernel - right ?
If so, I will post at linux-kernel.
If not, educate me :-)

Margit

Re: 2.5.70 Adaptec error messages

[Justin T. Gibbs]

> At 10:01 05.06.2003 -0600, Justin wrote:
>> Someone is fetching the device serial number without setting the
>> cmd underflow to 0.  The response length is variable, so telling
>> the adapter to fail the command when the returned result is less
>> than the allocated buffer size is counter productive.
> 
> Hmm, doesn't happen with 2.4.21rc + Justin's driver.
> Which, for me, would imply the kernel - right ?
> If so, I will post at linux-kernel.
> If not, educate me :-)
> 
> Margit

It is not the driver's fault unless the semantics of the underflow
field in the SCSI command structure have changed.

--
Justin

Re: 2.5.70 Adaptec error messages

[Margit Schubert-While]

At 10:40 05.06.2003 -0600, Justin wrote:
> > At 10:01 05.06.2003 -0600, Justin wrote:
> >> Someone is fetching the device serial number without setting the
> >> cmd underflow to 0.  The response length is variable, so telling
> >> the adapter to fail the command when the returned result is less
> >> than the allocated buffer size is counter productive.
> >
> > Hmm, doesn't happen with 2.4.21rc + Justin's driver.
> > Which, for me, would imply the kernel - right ?
> > If so, I will post at linux-kernel.
> > If not, educate me :-)
> >
> > Margit
>
>It is not the driver's fault unless the semantics of the underflow
>field in the SCSI command structure have changed.

I quite agree. I was trying to get at WHO is calling you with an incorrect
parameter. Is it somwhere else in 2.5 or is it possibly a user-space program
(although it doesn't appear with 2.4, so IMHO not likely) ??

Margit